The Zen Thread

[hobold]

Zen 5 announcement on AM5 a bit boring thus far. The same model line-up with 16, 12, 8, and 6 cores, now with Ryzen 9000 branding. An unexciting but solid 16% IPC improvement as per AMD's numbers, at the same clock speeds. But anything less than 16 cores now at lower, saner, TDP. No X3D models announced yet; presumably an ace up the sleeve waiting for whatever Intel will announce later this year.

Things are a bit more exciting with server and mobile Zen 5, where efficiency improvements directly lead to more performance and higher core counts.

Of the benchmarks AMD mentioned, gaming looked on par with Ryzen 7000X3D, so upcoming 9000X3D models are bound to be very nice, whenever they will eventually show up. The high outlier at >30% speedup was indeed an AVX-512 workload, as I had guessed earlier.

A new mainboard tier X870(E) with mandatory USB4 will come; mid-range seems to retain the B650 branding, but new models with USB4 seem to be waiting in the wings.

I am not disappointed like some other folks who were more hyped up. IMHO it is remarkable that double-digit IPC improvements are still possible even when transistors barely shrink anymore these days. Hopefully we'll learn more details about Zen 5's AVX-512 implementation soon. I have had some success toying with std::experimental::simd (a proposed C++ programming interface for using the various SIMD extensions of various processor families). It's looking good, so maybe towards the end of the year I'll be doing more SIMD programming on a Ryzen 9000 machine of my own.

 

[Paladin]

Yeah, honestly, if they could confidently say that they have eliminated a lot of potential security and stability issues, worked out the memory training and stability issues and worked to make motherboards both cheaper and more realiable with their chips and had just a decent increase in performance and efficiency, I would be raving about them.

 

[Anonymous Chicken]

Yeah I'm pretty impressed too. It seemed at one point that further IPC wasn't realistic, yet it keeps coming, without apparent drawbacks.

The mobile-first die is closing the gap with the server-first minimal product (i.e. 9600X).

 

[rodalpho]

Pretty good for a year, yeah. Wish they would launch the X3D versions simultaneously, or perhaps preannounce them for Q4 with pricing.

 

[IceStorm]

I'd want someone to do some API testing before saying Zen 5 is as fast as Zen 4 with vcache. vcache on Zen 3 makes a massive difference in how DX11 performs on AMD's GPU drivers. As it stands, I'd have to buy Zen 4 and Zen 5 to test it on my own, and I'm just not that invested in AMD's second-tier GPU ecosystem yet.

 

[mpat]

Note that the new mobile chips have 4 "full" Zen 5-cores and 8 smaller Zen 5c-cores.

 

[Anonymous Chicken]

Note that the new mobile chips have 4 "full" Zen 5-cores and 8 smaller Zen 5c-cores.

A brilliant move of course. There isn't power for 12 mega-turbo cores anyway. AMD has totally crushed Intel on this, if you ask me.

 

[mpat]

A brilliant move of course. There isn't power for 12 mega-turbo cores anyway. AMD has totally crushed Intel on this, if you ask me.

Depends on the inter-core latency. Presumably this has been done by making 3 clusters of 4 cores. Desktop Zen 5 has 8 or 16 cores, most likely still in clusters of 8 cores like Zen 3 and 4. If you remember the Zen 3 launch, use cases where you needed more than 4 cores on one task were improved by moving to one cluster of 8.

Other than that, I agree that it is likely more power-efficient.

 

[rodalpho]

I'd want someone to do some API testing before saying Zen 5 is as fast as Zen 4 with vcache.

The 7700X is basically as fast in games as the 5800X3D so I wouldn’t be surprised but I agree of course wait on reviews.

 

[Asral]

Zen 5 looks like a very nice upgrade for me. I'm curious to see what TDP 9800X3D will land on now that 9700X is a 65W CPU.

 

[Anonymous Chicken]

Presumably this has been done by making 3 clusters of 4 cores.

I don't see that splitting into 3 groups is realistic. At that point they need not have bothered adding cores.

 

[Aeonsim]

The doubled AVX-512 performance will likely be useful for some of my workloads. I assume they've upgraded the AVX-512 engines from the previous half-width (using 256b width I believe, so 2 cycles per operation) to full width? It will be interesting to see if they can run these at full clock speed. A very different direction of travel from Intel which appears to be pulling AVX-512 back from everything that isn't server focused (lunar-lake apparently removes the hardware rather than disabling it).

 

[Made in Hurry]

The 9700X is an interesting chip to me, but i am in no rush at all.

 

[IceStorm]

After ripping yet another AM4 CPU out of its socket when removing its heatsink, I'm kinda ready for AM5 to be good enough for me to switch.

 

[BigLan]

You don't try to twist it to loosen the thermal compound first instead of just trying to hulk it off?

 

[IceStorm]

It was completely warmed up, and it was a Noctua L12 in a Ghost S1. Not a whole lot of room to twist.

 

[redleader]

The doubled AVX-512 performance will likely be useful for some of my workloads. I assume they've upgraded the AVX-512 engines from the previous half-width (using 256b width I believe, so 2 cycles per operation) to full width? It will be interesting to see if they can run these at full clock speed. A very different direction of travel from Intel which appears to be pulling AVX-512 back from everything that isn't server focused (lunar-lake apparently removes the hardware rather than disabling it).

Pretty sure it was 1 cycle per operation throughput but it used both EUs so you couldn't dual issue.

Intel's new E cores now have 512 bit FMA, similar to Zen 4, which is hopefully a step towards finally unifying everything with AVX10. Maybe Zen 6 and Panther Lake finally fix this situation sometime next year and we can get past AVX classic, which will be at least 17 years old by the time both Intel and AMD support its replacement concurrently on consumer desktops (AVX2 launched 11 years ago today, although its just AVX1 with integer instructions).

 

[hobold]

Intel's new E cores now have 512 bit FMA, similar to Zen 4, which is hopefully a step towards finally unifying everything with AVX10.

At the current 256bit wide spec, AVX10 is more of a lowest common denominator rather than a unification. I hope std::experimental::simd will gradate out of experimental state sooner rather than later. That would partly unify SSE, AVX, AVX-512, NEON, SVE, AltiVec/VMX and whatever else is out there.

 

[redleader]

At the current 256bit wide spec, AVX10 is more of a lowest common denominator rather than a unification. I hope std::experimental::simd will gradate out of experimental state sooner rather than later. That would partly unify SSE, AVX, AVX-512, NEON, SVE, AltiVec/VMX and whatever else is out there.

AVX10 in the 256 bit flavor is still a massive improvement over AVX1/2, which really antiquated and essentially the lowest common denominator in that most operations are structured as two SSE operations run in parallel. Since all of these devices are going to have 512 bit wide vector units, if you access them using 2x256 AVX10 operations issued per cycle or 1x512 AVX10 operation per cycle is not so interesting. Key thing is just to get rid of the baggage and expose all the capabilities of those vector engines that aren't in AVX1/2 because mid-2000s CPUs didn't have the transistor budgets.

 

[hobold]

AVX10 in the 256 bit flavor is still a massive improvement over AVX1/2

Of course the capabilities of AVX-512, and thus AVX10, are on another planet. After all, AVX-512 was the first time that Intel aimed at a fully fledged, compiler targeted SIMD ISA in the tradition of DEC Alpha's "Tarantula" project, which in turn was based on all the experience gained from Cray's original vector supercomputers.

My "lowest common denominator" snide was aimed at Intel's goal of down-sizing AVX-512 enough to make it workable on their ex-Atom, now E-core, microarchitectures. I don't think Intel is doing the right thing, because AMD has already demonstrated by example that full AVX-512 support can be had with 256 bit wide hardware, and at very respectable performance, in Zen4. (And power consumption when running AVX-512 number crunching is not a problem on the Ryzen 7840/8840 laptops that I am toying with.)

Intel could have done the same and avoided the confusion of introducing yet another incompatible ISA / feature level. Especially considering that Intel still has to support AVX-512 on (some of) their datacenter silicon.

There are too many target platforms already, and programmer effort for tuning is already spread too thin. AVX10 does not help the situation. Pat Gelsinger wants "five nodes in four years" and "unquestioned technology leadership". Yet the actions of his company are telegraphing to the world that Intel will not have competitive transistor density, and cannot affort to implement AVX-512 in their client processors? I don't like it.

But yes, in the grand scheme of things, getting AVX10 is much, much better than being stuck on AVX2 forever.

 

[redleader]

My "lowest common denominator" snide was aimed at Intel's goal of down-sizing AVX-512 enough to make it workable on their ex-Atom, now E-core, microarchitectures. I don't think Intel is doing the right thing, because AMD has already demonstrated by example that full AVX-512 support can be had with 256 bit wide hardware, and at very respectable performance, in Zen4. (And power consumption when running AVX-512 number crunching is not a problem on the Ryzen 7840/8840 laptops that I am toying with.)

Zen 4's AVX512 is somewhat limited but still does one 512 bit FMA per cycle. With Crestmont also going to 512 bit wide FMA, it isn't clear what Intel is planning to do with AVX10. Or at least not to me.

 

[sakete]

Still don't know if I should upgrade from my 5950X. I want something that generates less heat overall. Maybe a 9800X3D once they announce that - I don't really need 16 cores.

 

[hobold]

I am tempted by Ryzen 9000 and will be eyeing the reviews closely. Playing with AVX-512 on Zen4 mobile has been successful and fun, and a 9950X would more than double most performance metrics of my aging 1950X machine. Let's see what prices and growing pains are going to be.

 

[Made in Hurry]

I do not really need a top of the line machine, but since AM4 computers still can be sold for quite a decent amount of money online these days, i listed my main rig online yesterday and i am going to keep that in my account and probably either build a new one before Christmas or buy a Mac.
I am curious if they have upgraded the integrated GPU on it though and what it can do as a mini-computer might be what i am tempted to get in such case.

 

[hobold]

I am curious if they have upgraded the integrated GPU on it

You mean Ryzen 9000 models? As far as I know the iGPU there is still minimal; no improvement over Ryzen 7000 desktop CPUs.

 

[Made in Hurry]

You mean Ryzen 9000 models? As far as I know the iGPU there is still minimal; no improvement over Ryzen 7000 desktop CPUs.

Thank you, yes i meant the 9000 series. I am sorry for not specifying that earlier.

 

[Demento]

The only reason to upgrade the desktop Ryzen iGPU is to be able to phase out support for that architecture in the drivers. It's there to provide basic - more basic than Intel - video and nothing more. I expect it will update the next time there's a major change in the Radeon architecture just to keep the drivers in sync.

 

[NervousEnergy]

The X3D line segregation seems... suboptimal if they want a big early sales splash. A lot of folks on X3D procs aren't going to upgrade until those chips are out, so I'm a bit surprised they don't lead with them. I'm still on a 5800X3D, and am definitely upgrading to X870E and whatever they're calling the 9K X3D versions (9800X3D?), so I'll be waiting.

 

[continuum]

The X3D line segregation seems... suboptimal if they want a big early sales splash. A lot of folks on X3D procs aren't going to upgrade until those chips are out,

Given how they've handled 3D V-cache parts in the past I don't think they care, they seem to be set on doing an initial release and then a subsequent 3D V-Cache part release.

 

[hobold]

they seem to be set on doing an initial release and then a subsequent 3D V-Cache part release.

Presumably the number of customers upgrading every single generation is small? I.e. AMD has little incentive to try and sell Ryzen 9000 to Ryzen 7000 owners, but rather directs their marketing efforts elsewhere.

 

[Demento]

I had bought the 7600 with the intent of shifting it to the Ivy Bridge machine once Win10 support ends and replacing it in my main PC with a 9800X3D. But I'm going to be part of a pretty small group there.

 

[Thegn]

I only bought into the 7000 series because my 1600x was getting long in the tooth, and for whatever reason my mobo wouldn't accept the BIOS update needed to get a 5000 series part. My plan is to sit on my config until the last generation of processors for my motherboard, get the best deal there, and sit on it until Zen 6.

 

[hobold]

The Zen 5 (Ryzen 9000) plot is thickening. Various benchmarks from six, eight, and twelve core models have leaked out, and they seem to confirm the ~16% uplift compared to the Zen 4 predecessors. IMHO it is looking like these chips are qualification samples rather than engineering samples. This would imply that final clock speed targets have been reached. In cases when power consumption is quoted with the benchmark score, it looks to be notably lower than previous gen models running the same benchmark.

All these information tidbits are unofficial rumors. But they seem to be lining up perfectly with what's been disclosed by AMD so far. It looks as though AMD will be able to maintain their momentum with Ryzen 9000.

(No sign of X3D models as of yet. Presumably AMD will maintain their record of executing their CPU plans reliably, so gamers might want to wait a bit longer for an 9800X3D or whatever it will be called.)

 

[continuum]

(No sign of X3D models as of yet. Presumably AMD will maintain their record of executing their CPU plans reliably, so gamers might want to wait a bit longer for an 9800X3D or whatever it will be called.)

Hopefully not too long a wait, but always hard to tell with rumors-- maybe as early as end of Q3...

AMD Ryzen 9000X3D "Zen 5" CPUs To Feature Same 3D V-Cache As Ryzen 7000X3D: 9950X & 9900X With 128 MB, 9800X3D With 96 MB L3

AMD's Ryzen 9000X3D "Zen 5" Desktop CPUs are expected to feature the same amount of 3D V-Cache as the existing Ryzen 7000X3D chips.

wccftech.com

 

[Drizzt321]

I thought I was going to upgrade to Zen 4, but the 5800X3D was FAR too cheap and performant to keep my X370 platform alive. I might just go for Zen 5, after the initial surge of X3D buying. Sees to provide a worthy gaming boost, and will let me keep the core system another 5-7 years, maybe with a CPU upgrade at some point in time with the last compatible CPU gen on the newer platform.

 

[hobold]

The thing I like most is the efficiency improvement. Because I am still hoping that AMD would bring more than 16 cores to AM5. It will probably take some prodding by Intel's sea of E-cores, and that might not happen. But it is good to know that AMD could hypothetically bring 8 fast + 16 compact Zen5 cores, and would not be totally crippled by power consumption - such a hybrid SKU wouldn't be pointless.

 

[Drizzt321]

I'll see what performance is, I'll probably stick with 8 cores, but the 9950X3D with 12 cores is a maybe for me. Depending on price. Definitely going 2x64GB though (some big VMs for work), and probably switching to a mATX with a new case too.

 

[IceStorm]

Making the same mismatched CCD vcache mistake as they did with the 7000 series. So sad.

 

[Drizzt321]

Oh wait, yeah. Hm. Well, it's far enough out I won't worry about it, I'll decide when I get to it.

 

[BigLan]

Zen 5 desktop release is set for July 31, but no prices yet - guess they don't want to give Intel time to react. The tech briefing of 15% IPC seems legit, and AMD is touting that the 9700x is faster than the 5800x3d for games.

More details over at anandtech, which feels like the last old-school media outlet that hasn't switched to video content. https://www.anandtech.com/show/2146...bile-strix-point-with-rdna-35-igpu-xdna-2-npu